Predicting cognitive load in immersive driving scenarios with a hybrid CNN-RNN model
URL: http://arxiv.org/abs/2408.06350v1
archive: archived pipeline: cataloged verified
Abstract
One debatable issue in traffic safety research is that cognitive load from sec-ondary tasks reduces primary task performance, such as driving. Although physiological signals have been extensively used in driving-related research to assess cognitive load, only a few studies have specifically focused on high cognitive load scenarios. Most existing studies tend to examine moderate or low levels of cognitive load In this study, we adopted an auditory version of the n-back task of three levels as a cognitively loading secondary task while driving in a driving simulator. During the simultaneous execution of driving and the n-back task, we recorded fNIRS, eye-tracking, and driving behavior data to predict cognitive load at three different levels. To the best of our knowledge, this combination of data sources has never been used before. Un-like most previous studies that utilize binary classification of cognitive load and driving in conditions without traffic, our study involved three levels of cognitive load, with drivers operating in normal traffic conditions under low visibility, specifically during nighttime and rainy weather. We proposed a hybrid neural network combining a 1D Convolutional Neural Network and a Recurrent Neural Network to predict cognitive load. Our experimental re-sults demonstrate that the proposed model, with fewer parameters, increases accuracy from 99.82% to 99.99% using physiological data, and from 87.26% to 92.02% using driving behavior data alone. This significant improvement highlights the effectiveness of our hybrid neural network in accurately pre-dicting cognitive load during driving under challenging conditions.
Summary
Khan et al. extended their fNIRS/eye-tracking/driving-behavior fusion approach to a hybrid 1D-CNN + RNN architecture for three-level cognitive-load classification under low-visibility (nighttime, rainy) simulated driving. Ten adults performed an auditory 0/1/2-back task while driving on a simulated freeway with traffic; the proposed model was compared against prior CNN-LSTM baselines. The hybrid CNN-RNN raised accuracy from 99.82% to 99.99% on physiological data and from 87.26% to 92.02% on driving-behavior data alone, while reducing parameter count. The authors argue physiological signals remain decisive for fine-grained workload discrimination but driving-behavior models can be improved with better architectures.
Key finding
A lighter hybrid 1D-CNN + RNN improved three-level cognitive-load classification to 99.99% with physiological signals and 92.02% with driving-behavior signals, surpassing prior CNN-LSTM baselines.
Methodology
Driving simulator study with 10 healthy adults (9M, 1F) performing three levels of an auditory n-back secondary task (0/1/2-back) during simulated nighttime/rainy freeway driving with traffic. Recorded fNIRS, eye-tracking (Pupil Core), and driving behavior. Trained a hybrid 1D-CNN + RNN classifier on physiological-only and driving-behavior-only feature sets; benchmarked against CNN-LSTM baselines on accuracy and parameter count.
Sample size: N=10 (9 male, 1 female)
Quality score: 5 / 5